[11:36] <bac> ahoy my yellow friends
[11:59] <gary_poster> hiya
[12:01] <gary_poster> nice trip report frankban.  about concurrent.futures you said "Guido doesn't like them...".  Did you mean concurrent.futures or futures.as_completed?
[12:05] <frankban> gary_poster: I mean callbacks
[12:05] <gary_poster> frankban, oh!  I see
[12:05] <gary_poster> interesting.  Wil look at that more closely later
[12:09] <gary_poster> bac benji frankban gmb, call in 1
[12:09] <gary_poster> @ http://tinyurl.com/yellowsquad
[12:11] <gary_poster> gmb you on the way?
[13:28] <gary_poster> benji, re tarmac automation: got a lot of pushback about bootstrap not sending the key properly (I didn't help anything by doing a classic "gary gets off on the wrong foot by calling a tomato a potato" mistake, where in this case it was calling bootstrap deploy).  However, hazmat suggested that we could set $HOME
[13:28] <gary_poster> and then generate .juju and .ssh
[13:29] <gary_poster> not a horrible idea
[13:29] <benji> interesting
[13:29] <benji> a hack, but still interesting
[13:31] <gary_poster> benji, meanwhile, on #juju other people are having interesting ideas
[13:32]  * benji looks
[13:32] <gary_poster> by from non-direct-juju people :-)
[13:33] <gary_poster> frankban, http://tinyurl.com/yellowsquad ?
[13:33] <frankban> gary_poster: joining
[13:37] <benji> The best solution for us to implement right away seems to be either "nothing" (i.e., keep it the way it works now with having to tweak your ssh config) or hacking $HOME for the duration of the test run
[13:38] <benji> hmm, but hacking $HOME may well not work if the user has some ssh config that needs to apply to all machines (like always bouncing the connection through a gatekeeper machine)
[13:39] <benji> I guess we could copy their .ssh/config to the new $HOME location and then append our shut-up-AWS config to the end, but it is starting to get really hacky
[14:23] <gary_poster> benji, agree, thought as much
[14:23] <gary_poster> gmb, call in 7?
[14:23] <gmb> gary_poster, Yep.
[14:23] <gary_poster> cool
[14:29] <gmb> gary_poster, Yellow hangout or elsewhere?
[14:30] <gary_poster> gmb, yellow hangout
[14:30] <gmb> kk
[14:45] <bac> gary_poster: create_lxc, when in install_lxc, had an --install-subunit option.  it looks like it got dropped in the move to initlxc.py.  that should be added back, no?
[14:46] <gary_poster> yes good idea: init-lxc should have that option
[15:08] <gmb> gary_poster, Lost you
[15:50] <bac> gary_poster: i'm making progress but might benefit from pairing if you're available after lunch
[15:52] <gary_poster> bac, ok, sounds good
[15:56] <bac> gary_poster: ok, i'm grabbing some food now
[15:56] <gary_poster> cool bac, will do the same in a few
[16:46] <frankban> gary_poster: what do you think about adding a check for repository, branch and checkout directories existence in init-repo? Exiting with an error if they already exist?
[16:59] <gary_poster> frankban, +1
[16:59] <frankban> cool
[16:59] <gary_poster> frankban, we will need to think pretty caefully about the developer story
[16:59] <gary_poster> frankban, so far it's been fairly neglected other than the big picture
[17:00] <gary_poster> that's the kind of change will be good for developers, I think
[17:03] <frankban> gary_poster: I agree. I am also inclined to rename settings.CHECKOUT_DIR to settings.LP_REPOSITORY_DIR. CHECKOUT_DIR seems confusing...
[17:05] <gary_poster> frankban, a lot of the names have been confusing since the refactoring.  I thought we fixed that one though--isn't that the name of the default checkout directory, rather than the path to the default repository
[17:05] <gary_poster> ?
[17:06] <frankban> gary_poster: in init-repo CHECKOUT_DIR is the default for --repository, and LP_CHECKOUT_NAME is the default for --checkout-name
[17:06] <gary_poster> frankban, heh, ok, yeah you are right :-)
[17:08] <benji> gary_poster: I managed to get the "originals" backup directory working, but I think we should discuss the security implications
[17:08] <gary_poster> great, and, uh-oh :-)
[17:09] <benji> :)
[17:10] <gary_poster> benji, I'll ping in just a few more
[17:10] <benji> k
[17:36] <gary_poster> benji, yellowsquad when you are ready
[18:47] <benji> gary_poster: I'm coming back
[19:28] <bac> gary_poster: ready at 3:30?
[19:28] <gary_poster> bac, yes
[19:47] <bac> benji: http://paste.ubuntu.com/1088608/  <- tb traceback
[19:48] <benji> bac: is this on a machine you've never used tb on before?
[19:48] <bac> benji: yes
[19:48] <bac> gary_poster: trying to get plugin installed
[19:48] <gary_poster> :-) k
[19:49] <bac> gah
[19:50] <benji> bac: hmm, I may not have the dependencies or PPA set up right then, it sounds like you need my patched version of python-vte (or whatever the package name is)
[19:50] <benji> bac: try sudo apt-get install python-vte
[20:43] <bac> gary_poster: you there?
[20:43] <gary_poster> bac, yes
[20:43] <gary_poster> bac, I'm here now
[21:14] <benji> bac: did you get tb happy?
[21:14] <bac> benji: yes,thanks.  sorry i didn't give you feedback
[21:15] <benji> bac: was it my suggestion that fixed it?  if so, I need to figure out why that wasn't automatically installed when you apt-get installed termbeamer
[21:17] <bac> yes, installing python-vte made it work
[21:18] <bac> benji: ^
[21:19] <bac> benji: http://paste.ubuntu.com/1088767/
[21:19] <bac> what are those 0s and nulls?